Frequently asked questions
Everything we get asked about Populate.ai — how it works, what's supported, privacy, and troubleshooting.
The basics
- Populate.ai is a browser-first tool that turns bills, invoices, receipts, photos, and dictated notes into clean, structured data you can export to CSV, XLS, or PDF. A 5-agent pipeline runs OCR, cleans the text, spell-checks, translates if needed, and scores confidence before showing you the result.
- Freelancers, accountants, small businesses, and anyone tired of typing line items from a paper receipt into a spreadsheet. If you process a handful of documents a week or hundreds a day, it works the same.
- Yes. The current version is free to use. We may introduce paid plans for very high-volume users later, but the core extractor will remain free.
- No. Open the site and start uploading — there's nothing to sign up for.
What is Populate.ai?+
Who is it for?+
Is it free?+
Do I need to create an account?+
Uploading documents
- PDF, PNG, JPG, and common audio formats (WebM, MP4/M4A, MP3, WAV, OGG, AAC) for dictation.
- Each file is capped at 15 MB. This is a free service and the cap keeps processing fast and reliable for everyone.
- Yes — drag and drop multiple files, or select several in the picker. They process sequentially through the 5-agent pipeline.
- Yes. Tap "Take photo" on a mobile device and your camera opens directly. Aim for good lighting and a flat surface for the best OCR.
- Tap "Dictate" to record yourself describing an expense — e.g. "Lunch with Maya, twenty-two euros, June 12, paid by card." Populate.ai transcribes and structures it just like a scanned receipt.
What file types can I upload?+
Is there a file size limit?+
Can I upload more than one file at a time?+
Can I take a photo with my phone?+
What is dictation?+
Privacy & security
- Your documents are processed only to extract your data and are never stored on our servers or retained after processing. Extraction is performed via a secure AI processing service; we keep no copy of your files. The resulting structured data is kept in your browser's local storage so you can revisit it.
- No. Your documents are not used to train any model.
- Yes — every file is checked in your browser before processing. We verify the file's actual contents match its declared type, block executables and archives, and scan PDFs for active-content tokens commonly used by malware.
- Documents you've already processed are kept in your browser's local storage so you can revisit them from "Recent documents." Clearing your browser data removes them.
Where do my documents go?+
Do you train models on my data?+
Is there an antivirus scan?+
What happens if I close the tab?+
Accuracy & the pipeline
- Each field gets a confidence score. Anything below 90% is automatically routed to the Review queue so you can verify it. In practice, clean PDFs score very high; photos of crumpled receipts vary.
- Extract (OCR / vision), Clean (normalize text), Spell-check (fix recognition errors), Translate (optional, into your preferred language), and Score (confidence + flagging for review).
- OCR handles the major Latin-script languages and many non-Latin scripts out of the box. The translate step can render the cleaned output in English or your preferred target language.
- Open the document from "Recent" or the "Review queue," click any field, and edit it. Corrections are saved instantly.
How accurate is the extraction?+
What are the 5 agents?+
What languages are supported?+
What if a field is wrong?+
Exporting & integrations
- CSV (universal spreadsheet), XLS (Excel), and a formatted PDF summary. CSV is the most flexible for downstream tools.
- Direct integrations aren't built in yet. For now, export CSV/XLS and import into your tool of choice. Tell us which integration you'd like — popular requests get prioritized.
What export formats are available?+
Can I push results into my accounting tool?+
Troubleshooting
- The file looked like an executable, an archive, or a PDF with active content. Re-export it as a plain PDF ("Print to PDF" works well) or save the image as PNG/JPG and try again.
- Re-scan with brighter, even lighting, lay the document flat, and avoid heavy shadows. For phone photos, hold the camera straight on (not angled).
- Refresh the page — your file isn't lost, processing resumes from your local queue. If it happens repeatedly, try a smaller file or a different format.
- Documents live in your browser. If you switch browsers or devices, or clear site data, they won't appear. We're considering opt-in sync — let us know if you'd use it.
My upload was blocked by the security scan — why?+
OCR confidence is low — what can I do?+
The page seems stuck on "processing."+
I don't see my old documents.+
Roadmap & feedback
- Email us or use the feedback link in the footer of any results page. Specific examples (a redacted screenshot, the field that was wrong) help us fix things fast.
- Better multi-page invoice handling, more export targets, opt-in cloud sync, and accounting tool integrations. The roadmap is shaped by what users ask for most.